Rev Austrian Econ (2008) 21:45–59
DOI 10.1007/s11138-007-0028-y
Mises’ democracy–dictatorship equivalence theorem:
A critique
Bryan Caplan
Published online: 14 November 2007
# Springer Science + Business Media, LLC 2007
Abstract Ludwig von Mises argues that public opinion, not the form of government, is the
ultimate determinant of policy. The implication is that, holding public opinion constant,
democracies and dictatorships will have the same policies—a result I call Mises’
Democracy–Dictatorship Equivalence Theorem. According to Mises, dictators have to
comply with public opinion or else they will be overthrown. I argue that he seriously
overestimates the power of revolution to discipline dictators. Mises was perceptive to note
that, in practice, “dictatorially imposed” policies are often democratically sustainable, but
he neglected several mechanisms—all more plausible than the threat of revolution—
capable of explaining this fact.
Keywords Democracy . Dictatorship . Public opinion . Revolution
JEL codes D72 . D74 . D70
1 Introduction
Ludwig von Mises vigorously defends democracy on utilitarian grounds. Unlike most
advocates of democracy, however, he denies that democracy causes better policies than
dictatorship. On Mises’ view, public opinion determines policy in democracies and
dictatorships alike.1 As he explains in Theory and History:
A statesman can succeed only insofar as his plans are adjusted to the climate of opinion
of his time, that is to the ideas that have got hold of his fellows’ minds. He can become
a leader only if he is prepared to guide people along the paths they want to walk and
toward the goal they want to attain. A statesman who antagonizes public opinion is
1
For a defense of Mises’ position and discussion of its precursors, see also Murray Rothbard’s introduction to
de la Boetie (1975: 9–42).
B. Caplan (*)
Department of Economics, Center for Study of Public Choice, and Mercatus Center,
George Mason University, Fairfax, VA 22030, USA
e-mail: [email protected]
46
B. Caplan
doomed to failure. No matter whether he is an autocrat or an officer of a democracy, the
politician must give the people what they wish to get, very much as a businessman must
supply the customers with the things they wish to acquire (1985a: 187).
The root cause of bad policies is not, therefore, the form of government but the quality
of public opinion. The advantage of democracy over dictatorship, for Mises, is simply that
it reaches the inevitable political outcome with minimal bloodshed:
Civil war and revolution are the means by which the discontented majorities
overthrow rulers and methods of government which do not suit them. For the sake
of domestic peace liberalism aims at democratic government. Democracy is therefore
not a revolutionary institution. On the contrary, it is the very means of preventing
revolutions and civil wars. It provides a method for the peaceful adjustment of
government to the will of the majority (1998a: 150).
One could even say that Mises sees democracy much as John Stuart Mill sees money: “It
is a machine for doing quickly and commodiously, what would be done, though less
quickly and commodiously, without it…” (Mill 2006: 506). The upshot, explored at length
by Mises’ student Murray Rothbard, is that political thinkers have to explain why every
government, even the most despotic, is popular—a problem that Rothbard dubs “the
mystery of civil obedience” (de la Boetie 1975: 13).
When governments adopt policies they disapprove of, many thinkers argue that public
opinion has been ignored. Mises will have none of this. Not just Roosevelt but Hitler
himself enjoyed popular support:
It is a fact that the policy of the New Deal has been supported by the voters. Nor is
there any doubt that this policy will be entirely abandoned if the voters withdraw their
favor from it. The United States is still a democracy. The Constitution is still intact.
Elections are still free. The voters do not cast their ballot under duress. It is therefore
not correct to say that the bureaucratic system carried its victory by unconstitutional
and undemocratic methods (1983a: 5).
It is probable that today about 80 per cent of all German-speaking Europeans are
Nazis. If we leave out the Jews, the Austrians, and the German-speaking Swiss, we
might say that more than 90 per cent of the Germans support Hitler’s fight for world
hegemony (1985b: 243–244).
For political theorists who assume that voters are rational, Mises’ position is hard to
accept. Many dictators choose policies that lead to poverty and war. How could such
policies be popular? Mises’ answer is that people suffer from systematically biased beliefs
about economics, foreign policy, and more (Caplan and Stringham 2005). They mistakenly
believe that counter-productive policies are the path to prosperity:
Democracy guarantees a system of government in accordance with the wishes and
plans of the majority. But it cannot prevent majorities from falling victim to erroneous
ideas and from adopting inappropriate policies which not only fail to realize the ends
aimed at but result in disaster (1998a, p.193).
In fact, Mises maintains that dictatorship itself is often popular. Many are willing to
ignore common sense and put their faith in a charismatic dictator:
It cannot be denied that dictatorship, interventionism, and socialism are extremely
popular today. No argument of logic can weaken this popularity (1998b: 85).
Mises’ democracy–dictatorship equivalence theorem: A critique
47
But what precisely is the mechanism that, holding public opinion constant, leads
democracy and dictatorship to converge on equivalent policies? Does Mises recognize any
loopholes in what we may call his Democracy–Dictatorship Equivalence Theorem? More
importantly, is his analysis correct? Is it really true that—collateral damage aside—we should
not expect replacing dictatorship with democracy to lead to systematic changes in policy?
In an earlier paper (Caplan and Stringham 2005), my co-author Edward Stringham and I
argued that Mises was a political economist ahead of his time, whose writings offer
promising directions for new research. The current paper pursues one of these directions. It
explores Mises’ view that the threat of revolution forces dictators to choose the same
policies that would have prevailed under democracy. It then argues that Mises overlooked
several compelling objections to his account. Despite these weaknesses, however, Mises’
conclusion is better than his arguments. Although dictators are unlikely to be overthrown
for defying public opinion, there are several alternative mechanisms that help make
dictators’ policies popular.
2 Mises’ democracy–dictatorship equivalence theorem
Does Mises really believe that, holding public opinion fixed, democracy and dictatorship
lead to the same policies? There are definitely times where he appears to claim the opposite.
For example, in a discussion of Austrian business cycle theory, Mises remarks that:
[I]n a socialist economy it is only the government’s value judgments that count, and the
people are deprived of any means of making their own value judgments prevail. A
dictator does not bother about whether or not the masses approve of his decision
concerning how much to devote for current consumption and how much for additional
investment. If the dictator invests more and thus curtails the means available for current
consumption, the people must eat less and hold their tongues. No crisis emerges because
the subjects have no opportunity to utter their dissatisfaction (1998a: 565–566).
But this is not necessarily inconsistent with Mises’ broader perspective. A dictator could
ignore his subjects’ preferences because they want him to. A majority might favor car
inspections, although no one likes to get one, or favor Prohibition, although most people want
to drink.
In any case, when Mises specifically compares democracy to dictatorship or analyzes the
power of public opinion, he almost invariably takes the position I ascribe to him.2 He insists
that, regardless of the form of government, public opinion determines economic policy:
What determines the course of a nation’s economic policies is always the economic
ideas held by public opinion. No government, whether democratic or dictatorial, can
free itself from the sway of the generally accepted ideology (1998a: 850).
2
The one clear exception I have been able to find appears in Nation, State, and Economy, where Mises
states:
In Russia socialism certainly is not a movement of the immense majority. That it claims to be a
movement in the interest of the immense majority is nothing special; all movements have claimed that. It
is certain that the rule of the Bolsheviks in Russia rests just as much on possession of the government
apparatus as the rule of the Romanovs once did. A democratic Russia would not be Bolshevik (1983b:
204; emphasis added).
48
B. Caplan
However, economic policy is only a special case of a more general rule, which Mises
dubs “the supremacy of public opinion”:
The supremacy of public opinion determines not only the singular role that economics
occupies in the complex of thought and knowledge. It determines the whole process of
human history (1998a: 863).
Mises derides “great man” theories. Even seemingly all-powerful dictators like Hitler
basically had to do what their countrymen wanted them to do:
Simplified accounts of history, adapted to the capacity of people slow of
comprehension, have presented history as a product of the feats of great men. The
older Hohenzollern made Prussia, Bismarck made the Second Reich, William II ruined
it, Hitler made and ruined the Third Reich. No serious historian ever shared in such
nonsense… Every man, whether great or small, lives and acts within the frame of his
age’s historical circumstances. These circumstances are determined by all the ideas
and events of the preceding ages as well as by those of his own age. The Titan may
outweigh each of his contemporaries; he is no match for the united forces of the
dwarfs (1985a: 186–187).
Holding public opinion constant, parliamentary safeguards are unable to check the
excesses of even the most notorious dictators because their “excesses” are a byproduct of
the statist policies the majority wants:
It is not an accident that socialist countries are ruled in a dictatorial way.
Totalitarianism and government by the people are irreconcilable. Things in Germany
and Russia would not be different if Hitler and Stalin were to submit all their decrees
to the decision of their “parliaments.” Under government control of business
parliaments cannot be anything else than assemblies of “yes” men (1983a: 8).
What exactly is the mechanism that harmonizes public opinion and policy? In markets,
Mises repeatedly argues that entrepreneurs quickly go out of business unless they satisfy
their customers: “If they fail to produce in the cheapest and best possible way the
commodities which the consumers are asking for most urgently, they suffer losses and are
finally eliminated from their entrepreneurial position” (1980: 108). Mises claims that a
similarly strong selection mechanism weeds out politicians who fail to satisfy public
opinion3:
No matter what the constitution of the country, governments always have to pursue that
policy which is deemed right and beneficial by public opinion. Were they to attempt to
Mises does however maintain that, in a sense, the market is “more democratic” than democracy. To
describe the market “as a democracy in which every penny gives a right to cast a ballot” is overly generous to
democracy:
3
It would be more correct to say that a democratic constitution is a scheme to assign to the citizens in
the conduct of government the same supremacy the market economy gives them in their capacity as
consumers. However, the comparison is imperfect. In the political democracy only the votes cast for
the majority candidate or the majority plan are effective in shaping the course of affairs. The votes
polled by the minority do not directly influence policies. But on the market no vote is cast in vain
(1998a: 271).
Mises’ democracy–dictatorship equivalence theorem: A critique
49
stand up against the prevailing doctrines they would very soon lose their positions to
men willing to conform to the demands of the man in the street (1998b: xii).
According to Mises, all governments have some version of this selection mechanism.
Under democracy, the selection mechanism is elections. Politicians who act contrary to
public opinion are unable to get enough votes to stay in office. Under dictatorship, in
contrast, the selection mechanism is revolution:
The difference between democracy and despotism does not affect the final outcome. It
refers only to the method by which the adjustment of the system of government to the
ideology held by public opinion is brought about. Unpopular autocrats can only be
dethroned by revolutionary upheavals, while unpopular democratic rulers are
peacefully ousted in the next election (1998a: 863).
Why would these selection mechanisms yield the same outcomes? First, in both
elections and violent conflicts, people join the side they prefer to win.4 Second, in both
elections and violent conflicts, the side with more people wins:
The ultimate victor to emerge from such conflicts will be the faction strongest in
number. In the long run, a minority—even if it is composed of the most capable and
energetic—cannot succeed in resisting the majority (1978: 50).
The tyrant must have a retinue of partisans who obey his orders of their own accord.
Their spontaneous obedience provides him with the apparatus he needs for the
conquest of other people. Whether or not he succeeds in making his sway last depends
on the numerical relation of the two groups, those who support him voluntarily and
those whom he beats into submission. Though a tyrant may temporarily rule through a
minority if this minority is armed and the majority is not, in the long run a minority
cannot keep the majority in subservience. The oppressed will rise in rebellion and cast
off the yoke of tyranny (1998a: 189).
One might reply that numerical superiority is only one factor out of many that determine
the victor in a violent conflict. But for Mises, numbers are fundamental. He does not
hesitate to infer that restricting democracy to promote liberty is futile:
If judicious men see their nation, or all the nations of the world, on the road to
destruction, and if they find it impossible to induce their fellow citizens to heed their
counsel, they may be inclined to think it only fair and just to resort to any means
whatever, in so far as it is feasible and will lead to the desired goal, in order to save
everyone from disaster. Then the idea of a dictatorship of the elite, of a government by
the minority maintained in power by force and ruling in the interests of all, may arise
and find supporters. But force is never a means of overcoming these difficulties. The
tyranny of a minority can never endure unless it succeeds in convincing the majority
of the necessity or, at any rate, of the utility, of its rule. But then the minority no
longer needs force to maintain itself in power (1978: 45).
4
In this discussion, Mises appears to be making empirical, not praxeological, claims about human behavior.
On praxeological grounds, Mises could justifiably claim that people join the side they prefer to join; he could
not, however, make the stronger claim that people join the side they prefer to win. After all, someone could
join the side he prefers to lose because it pays better or because he wants to be in the winning side.
50
B. Caplan
In fact, Mises maintains that resorting to violence is counter-productive because it
further alienates public opinion:
The suppression of all opposition by sheer violence is a most unsuitable way to win
adherents to one’s cause. Resort to naked force—that is, without justification in terms
of intellectual arguments accepted by public opinion—merely gains new friends for
those whom one is thereby trying to combat. In a battle between force and an idea, the
latter always prevails (1978: 50).
The one qualification that Mises grants is that public opinion is sure to prevail only “in the
long run”:
Only a group that can count on the consent of the governed can establish a lasting
regime. Whoever wants to see the world governed according to his own ideas must
strive for dominion over men’s minds. It is impossible, in the long run, to subject men
against their will to a regime that they reject (1978: 46).
Whatever the system of government may be, there cannot be any question of ruling a
nation lastingly on the ground of doctrines at variance with public opinion. In the end
the philosophy of the majority prevails. In the long run there cannot be any such thing
as an unpopular system of government (1998a: 863).
Mises admits that a minority can temporarily rule over a majority through superior
military prowess. But unless the minority persuades the majority of the legitimacy of its
rule, the political order is bound to change:
Minorities too can sometimes conquer by means of superior military skill and can thus
establish minority rule. But such an order of things cannot endure… All victorious
minorities who have established a lasting system of government have made their sway
durable by means of a belated ideological ascendancy. They have legitimized their
own supremacy either by submitting to the ideologies of the defeated or by
transforming them. Where neither of these two things took place, the oppressed many
dispossessed the oppressing few either by open rebellion or through the silent but
steadfast operation of ideological forces (1981a: 190).
Technological superiority is not a durable solution to this dilemma:
Victorious minorities sometimes owe their success to their technological superiority.
This does not alter the case. In the long run it is impossible to withhold the better arms
from the members of the majority. Not the equipment of their armed forces, but
ideological factors safeguarded the British in India (1981a: 191).
Of course, if the “long run” could last centuries, Mises’ claim would be vacuous rather
than bold. But he does not rely on this qualification to insulate his position from counterexamples. His “long-run” seems to be no more than a few years. In Mises’ analysis of the
Russian Revolution, for example, he freely admits that the majority of Russians bitterly
opposed Lenin’s agricultural policies:
[W]hen the Bolsheviks seized control in Russia, they were a small minority, and their
program found scant support among the great masses of their countrymen. For the
peasantry, who constitute the bulk of the Russian people, would have nothing to do
with the Bolshevik policy of farm collectivization. What they wanted was the division
of the land among the “landed poverty,” as the Bolsheviks call this part of the
population (1978: 45–46).
Mises’ democracy–dictatorship equivalence theorem: A critique
51
Despite his ruthlessness, Lenin soon had to give the peasants what they wanted:
In order to remain in power, Lenin and Trotzky not only accepted this agrarian reform,
but even made it a part of their own program, which they undertook to defend against
all attacks, domestic and foreign. Only thus were the Bolsheviks able to win the
confidence of the great mass of the Russian people. Since they adopted this policy of
land distribution, the Bolsheviks rule no longer against the will of the great mass of
the people, but with their consent and support. There were only two possible
alternatives open to them: either their program or their control of the government had
to be sacrificed. They chose the first and remained in power (1978: 46).
Mises strangely glosses over the Bolsheviks’ harsh treatment of the peasantry during the
War Communism period (Pipes 1990). Lenin’s practice of requisitioning to the point of
mass starvation was at least as unpopular as collectivization would have been. Nevertheless,
Mises could easily reply that Lenin defied public opinion for less than 4 years before—
facing the collapse of his regime—he had to adopt the New Economic Policy. If Lenin
could not impose unpopular policies with terror for more a few years, who could?
3 Flaws in the theorem
The answer, as any student of Russian history can attest, was Stalin. Mises’ analysis of
Bolshevik agricultural policy was published in 1927. One year later, Stalin began
collectivizing Soviet agriculture (Werth 1999; Conquest 1986). By 1933, despite mass
starvation and millions of deaths, private agriculture was virtually eliminated. There is no
reason to think that Soviet public opinion became more favorable to collectivization during
the NEP era. All historical evidence indicates that Stalin’s policy was exceedingly
unpopular. Most Soviet citizens were still farmers, and nothing during the 7 years from
1921–1928 convinced them that the government should seize their land. The hostility of
Soviet opinion toward collectivization may have moderated with passing decades,5 but
Stalin’s policies cannot be explained away as a short-run aberration.
It would be hard to find a more compelling counter-example to Mises’ view that even
dictators have to bow before public opinion. Plainly, there is some flaw in his reasoning. A
weaker version of Mises’ position—that dictators rarely stray far from public opinion—may
well be correct. However, at minimum, the mechanism that Mises appeals to must be less
reliable than he claims.6
3.1 The paradox of revolution
On a theoretical level, the deepest problem with Mises’ Democracy–Dictatorship Equivalence
Theorem is that he conflates mere preference for outcome X with willingness to fight to get X:
Whatever the system of government may be, the foundation upon which it is built and
rests is always the opinion of those ruled that to obey and to be loyal to this
5
According to Malia (1994: 284), collective farming remained extremely unpopular in 1941, when there
were high hopes that the German invaders would restore private property: “As the Germans advanced into
the Baltic states, Bielrussia, and Ukraine, civilians often welcomed them as liberators….In particular, they
hoped for the dissolution of the kolkhozes and the reopening of the churches….”
6
For a broader discussion of how dictators manage to stay in power while deviating from citizen preferences,
see Wintrobe (1998).
52
B. Caplan
government better serves their own interests than insurrection and the establishment of
another regime. The majority has the power to do away with an unpopular government
and uses this power whenever it becomes convinced that its own welfare requires it
(1998a: 149–150).
However, this violates methodological individualism. An individual who prefers
outcome X may still decide, out of prudence, to do nothing on behalf of X, or even
struggle against it. Just because you want Stalin to die does not imply that it is prudent to
try to assassinate him. Indeed, just because you want Stalin to die does not imply that it is
imprudent to be his faithful bodyguard. Perhaps Stalin pays well, and the probability of
successfully assassinating him is low.
But the problem is not just that people may be unwilling to fight for the political
outcomes they prefer. There are strong reasons to expect people to be unwilling to fight for
the political outcomes they prefer. After all, one person has little effect on the probability
that the revolution succeeds, and unsuccessful rebels face harsh punishments. As Gordon
Tullock explains:
Suppose we have our common citizen in a state with an oppressive and inefficient
government. Suppose also that we have a group of noble people who propose to
overturn this government and set up a virtuous and efficient government. Assume that
some individual believes their promises… Should he join a popular uprising? The
answer in general is no, and the reason is the simple old-fashioned economic argument
of a public good.
The individual will have very little effect on the outcome….It is likely that any
individual joining in the revolution as opposed to remaining at home and cheering will
normally make only a tiny change in the likelihood of success, perhaps improving the
likelihood of success from .53278 to .53279. On the other hand, his chances of being
injured are greatly increased. If he compares the benefit that he will receive from
participating in the revolution with its cost to him, he’s almost certain to find the
payoff is negative (1987: 63–64).
For Tullock, quiescence in the face of tyranny is another example of the Prisoners’
Dilemma, where individual selfishness leads to collective misfortune.7 Indeed, he argues that,
for practical purposes, the quality of government policy is irrelevant to the decision to rebel:
[I]ndividuals will ignore the public goods aspects of the revolution in deciding
whether to participate and on which side to participate. The important variables are the
rewards and punishments offered by the two sides and the risk of injury during the
fighting (1974: 39).
The counter-intuitive implication is that a dictator can stay in power although a majority
of his subjects—or even all of his subjects—prefer change. If everyone realizes that the
private cost of changing things exceeds the private benefit, then no one rebels. Tullock
therefore has a straightforward solution to Rothbard’s “mystery of civil obedience” (de la
Boetie 1975): Individuals obey because if they do not, they are highly unlikely to make a
difference and quite likely to get killed.
7
If successful revolutionaries punish free riders after victory, the strategic problem is more complex. The
prudent strategy is then to support whichever side is expected to win, making this a coordination game rather
than a Prisoners’ Dilemma (Tullock 1974).
Mises’ democracy–dictatorship equivalence theorem: A critique
53
The gap between Mises and Tullock could hardly be greater. Mises expects revolution to
break out once the majority decides it would be better off under a new regime.8 Tullock
expects revolutions to be non-existent as long as the dictator retains the loyalty of the
military and does not hesitate to use force. For Tullock, the puzzle is not that revolutions are
so rare, but that they happen at all.
Tullock addresses this puzzle—which he calls “the paradox of revolution”—in two main
ways.9 First, he argues that most alleged “popular” uprisings against tyranny were actually
fights between rival elites: “I don’t want to swear that there have been no cases in history in
which the people have risen and disposed of a tyrannical leader, but I have never come
across a clear-cut case”10 (1987: 53). Second, he admits that not everyone is as selfish as his
analysis assumes11:
There are undeniably individuals with strong public interest who are willing to take
great risks or sacrifice their lives to benefit other people. This is particularly true if
religion can be brought in….It turns out that such people are fairly rare, however
(1987: 66).
Despite these concessions, Tullock’s analysis casts serious doubt on how closely the
threat of revolution constrains dictatorships. On Mises’ account, people rebel out of
perceived self-interest. On Tullock’s account, people rebel only after they decide that their
cause is more important than their self-interest. For Mises, economic law predicts that
revolutions will follow unpopular policies.12 For Tullock, economic law predicts political
stability regardless of the popularity of policy; to get a revolution, you need a loophole—
like rebels willing to die or leaders unwilling to kill.
One might reply that fear would not deter a majority from rebelling because the rebels
could expect to win. In fact, if the majority is going to win anyway, it could use sticks and
8
A slight variation on Mises’ argument is that dictators who deviate from public opinion will be overthrown
not by revolution but by a coup d’etat. There is some evidence that coups are more likely after recessions, but
poverty per se does not seem to increase the probability of a coup (Galetovich and Sanhueza 2000). For a
detailed analysis of the economics of the coup, see Tullock (1974).
9
For a praxeological critique of Tullock’s paradox of revolution, see Kurrild-Klitgaard (1997).
Mises, in contrast, tells us that “History provides an abundance of striking examples to show that, in the
long run, even the most ruthless policy of repression does not suffice to maintain a government in power.”
(1978:45).
10
11
For further discussion of this point, see Hummel (1990).
Although Murray Rothbard officially supports Mises’ position (de la Boetie 1975), he also undercuts it. In
Power and Market (1977: 196), Rothbard pointedly denies that “a democratic election will yield the same
result as would have occurred if the majority had had to battle the minority in violent combat.” Among other
reasons, democracy ignores “inequality of interest or intensity of belief”:
12
Thus, 60% of the population may oppose a certain policy, or political party, while only 40% favor it. In a
democracy, the latter policy or party will be defeated. But suppose that the bulk of the 40% are passionate
enthusiasts for the measure or candidate, while the bulk of the 60% majority have only slight interest in
the entire affair. In the absence of democracy, far more of the passionate 40% would have been willing to
engage in a combat test than would the apathetic 60%. (1977: 197)
Note that Rothbard accepts the Tullockian premise that people often fail to fight on behalf of the policies they
prefer. The difference is that Rothbard blames apathy, whereas Tullock focuses on fear, along with
individuals’ low probability of tipping the scales.
54
B. Caplan
carrots to short-circuit resistance and free-riding alike. Once again, however, this argument
violates methodological individualism. Even if every potential rebel knows that the
majority is on his side, someone still has to act first (Kuran 1995). A prudent dictatorship
responds by harshly punishing first movers. By analogy, even if a majority agreed that
Esperanto was the world’s best language, people might not adopt it. After all, someone has
to start speaking Esperanto first. The costs of non-conformity may be enough to discourage
any individual from breaking with the herd (Schelling 1978).
3.2 Switching costs
Setting aside the paradox of revolution, there is another important flaw in Mises’
Equivalence Theorem: It ignores switching costs13 (Shapiro and Varian 1999; Klemperer
1995). Switching governments with revolution is deadly and destructive. A dictator can
strategically exploit this switching cost to flout public opinion.
To see this point, suppose the majority is certain that it can overthrow the dictator but
only after a bloody struggle. Suppose further that people are willing to join a revolution if
they believe they will be better off, on balance, after a regime change. A dictator could still
deviate somewhat from majority preferences. How? By making sure that the perceived cost
of his deviations from majority preferences is less than the perceived cost of a revolution.
On a theoretical level, this point holds for any status quo. You will not change your
Internet service provider just because a competitor offers you a better deal than what you
already have. You will only accept the competitor’s offer if the extra value exceeds the
switching costs. But on a practical level, the switching costs for overthrowing a dictatorship
are likely to dwarf the switching costs for Internet service providers. Suppose citizens know
that they have a 10% chance of dying in the course of deposing their ruler. There would be
no revolution unless most people felt that the ruler’s deviation from public opinion was
worse than a 10% chance of death. That is a massive switching cost—which translates into
a lot of slack for the dictator.
A further implication of the switching costs argument is that, contrary to Mises, there is
a strong reason to expect democracies to be closer to public opinion than dictatorships. As
Mises himself points out, it is much cheaper to change governments with elections rather
violence. Democratically changing a government is a minor inconvenience; violently
changing a government is life-threatening. The upshot: The majority will vote against the
status quo once the deviation from public opinion seems worse than the minor
inconvenience of a change in administration. Switching costs give elected governments a
little slack, but nothing compared to the slack enjoyed by the typical dictator.
13
This omission is striking because Mises was well aware of the importance of switching costs for capital
investment decisions:
A new machine, more efficient than those used previously, is constructed. Whether or not the plants
equipped with the old, less efficient machines will discard them in spite of the fact that they are still
utilizable and replace them by the new model depends on the degree of the new machine’s superiority.
Only if this superiority is great enough to compensate for the additional expenditure required, is the
scrapping of the old equipment economically sound. (1998a: 507–8)
I thank David Gordon for reminding me of this point.
Mises’ democracy–dictatorship equivalence theorem: A critique
55
4 Salvaging the theorem
Due to collective action problems, people do not violently rebel against dictatorships just
because they think they would be better off with a new leader. Even if they did, the
switching costs associated with violent rebellion leave dictators with a great deal of political
slack. The mechanism behind the Democracy–Dictatorship Equivalence Theorem is not
strong enough to move the weight Mises places upon it.
It would be a mistake to counter that Mises correctly predicts the finding that, on
average, democracies and dictatorships have very similar policies (Mulligan, Gil, and Salai-Martin 2004; for a contrary evidence, see Deacon and Saha 2006) and growth rates
(Przeworski et al 2000; Barro 1997). Mises’ position only implies that, holding public
opinion constant, moving from dictatorship to democracy will not change policy. It does not
imply that the average democracy will have the same policies as the average dictatorship.
After all, public opinion in democracies could easily differ from public opinion in
dictatorships. In fact, on Mises’ view, we should expect such a difference: If public opinion
is wrong-headed enough to favor dictatorship, it is likely to favor other foolish policies as
well.
Still, there is one striking fact that Mises’ theory can readily explain, and Tullock’s
cannot: When dictatorship gives way to democracy, policy changes are often small.
Granted, there is intriguing evidence that the variance of growth increases after dictators’
unexpected deaths (Jones and Olken 2005) and that, controlling for income and other
variables, democracies have better education and health outcomes than dictatorships
(Deacon and Saha 2006). However, it is still remarkable how often policies that were
blamed on a dictator’s whims turned out to be democratically sustainable—just as Mises
would predict.
For example, there has been a strong populist backlash against free-market policies in
the former Soviet bloc, and politicians with strong ties to the old regime, not to mention
unrepentant Communists, have done well in democratic elections (Pipes 2004; Fessenden
2002; Gryzmała-Busse 2002; Lewis 2001; Colton 2000). Similarly, after Pinochet
relinquished power in Chile, democratically elected politicians made little effort to return
to the socialist policies of Allende (Velasco 1994). And much to the disappointment of
democratic optimists, free elections in Palestine did not give power to politicians eager to
make peace with Israel (Shikaki 2004).
If, as I have argued against Mises, dictators have considerable political slack, why don’t
they take greater advantage of it? There are at least three mechanisms that could account for
the similarity between democratically chosen and dictatorially imposed policies.
The stationary bandit model Among economists, Mancur Olson’s “stationary bandit”
model is probably the most popular explanation for why dictators would want to mimic
democratic policies (Höijer 2004; Olson 2000; McGuire and Olson 1996; for a more
circumspect version of the argument, see Grossman 2000). As long as a dictator expects to
remain in power for a long time, he has an incentive to do precisely what most of his
citizens want: make the nation prosperous.14 The dictator is like a majority shareholder in a
14
Of course, if economically inefficient policies are actually popular (Caplan 2007; Caplan and Stringham
2005), then democratic and dictatorial policies will not converge. Wealth-maximizing stationary bandits will
defy public opinion and choose better policies than democracies would.
56
B. Caplan
corporation. Neither can be fired for doing a bad job. But the better the leaders do, the more
valuable their “shares” become. As McGuire and Olson explain:
[T]here is a hidden hand that leads encompassing and stable interests with
unquestioned coercive power to act, to a significant and surprising degree, in the
interests of the entire society including those who are subject to their power. The
outcome from stationary banditry is not nearly as bad as might have been supposed
(1996: 94).
There are however two serious weaknesses in this story. First, a ruler’s expected tenure
could easily go down as his nation’s prosperity goes up. A more prosperous nation is likely
to have more contact with the outside world, undermining not only the loyalty of the public,
but more importantly, the loyalty of the military and other elites. Second, the ruler of even a
poor country is likely to be so rich that his marginal utility of additional wealth is low.
Other motivations—like ideological purity or ordinary vanity—could easily take precedence over dictators’ desire for greater riches.
Shared preferences A simpler explanation for why dictatorial and democratic rule would
produce similar policies is that dictators’ preferences are usually similar to those of the people
they rule. They could deviate a lot more from public opinion and remain in power. Because
they largely agree with public opinion, however, they have little motivation to deviate. After
all, most dictators emerge out of the intellectual and political culture of the nations they rule.
Given the human impulse to conformity (Pinker 2002; Mosca 1939), we should expect a
dictator to share prevailing values and assumptions. This is especially plausible for longstanding dictatorships, like hereditary monarchies. Yet, even avowedly revolutionary
movements often fall back on familiar national traditions. Richard Pipes (1974) discusses
the czarist roots of Russian Communist rule at length; others (Fu 1993; Salisburg 1992) note
the continuities between Deng, Mao, and the Chinese emperors who preceded them.
Status quo bias But perhaps reverse causation is the simplest way to harmonize policy and
public opinion. There is considerable evidence (Samuelson and Zeckhauser 1988) that
human beings are subject to what psychologists call status quo bias, a natural human
tendency to prefer what we are used to.15 If Mises is correct that the masses are “dull and
mentally inert” (1985b: 123), status quo bias is precisely what we should expect.16 What
moral heuristic requires less effort to apply than “whatever is, is right”? Thus, if people
have had a Social Security program for as long as they can remember, they are likely to
believe that this program is a good idea. If people have been ruled by a dictator for as long
15
A slight variation on this mechanism is that people tend to believe what their governments tell them.
Interestingly enough, status quo bias is one of the pillars of Rothbard’s solution to his “mystery of civil
obedience.” The main difference is that Rothbard relies on sixteenth-century French political writer Etienne
de la Boetie, instead of modern psychologists; and to be fair, it is hard to improve upon La Boetie’s account:
16
[A]ll those things to which he is trained and accustomed seem natural to man… Thus custom becomes
the first reason for voluntary servitude… [M]en will grow accustomed to the idea that they have
always been in subjection, that their fathers lived in the same way; they will think they are obliged to
suffer this evil, and will persuade themselves by example and imitation of others, finally investing
those who order them around with proprietary rights, based on the idea that it has always been that
way (de la Boetie 1975: 64–65).
Mises’ democracy–dictatorship equivalence theorem: A critique
57
as they can remember, they are likely to believe that dictatorship is a good idea—or at least
better than any alternative. One might expect people who endured life under Communism
to be more pro-market than those who did not, but the opposite seems to be true: Alesina
and Fuchs-Schündeln (2005) actually find that East Germans are markedly more antimarket than West Germans, even controlling for income.
All this suggests that policy could drive public opinion, rather than the other way around.
With sufficiently strong status quo bias, public opinion and policy will match under every
form of government, just as Mises claimed. The difference, of course, is the mechanism.
Instead of dictators yielding to public opinion to preempt violent overthrow, perhaps dictators
do as they please, and—within broad limits—public opinion accepts their decisions on faith.
5 Conclusion
Mises questions the widespread view that dictatorship is less responsive to public opinion than
democracy. Maybe it is less responsive in the short run, he admits; but in the long run, all
dictatorship is majority dictatorship. The public can and will overthrow any government that
deviates from public opinion. “People power” intimidates most dictatorships and topples the rest.
There is little reason to think that this mechanism works as well as Mises claims. As
Tullock lucidly explains, joining a rebellion against even the worst government is an act of
charity, not an expression of self-interest. For Mises to claim that people will rebel as soon
as they realize that they would be better off under another government is simply mistaken.
Furthermore, even if Tullock were wrong, Mises would still seriously underestimate the
slack that the typical dictator enjoys. Ignoring all free-rider problems, we should not expect
people to rebel merely because they will be better off after the rebellion succeeds. Instead,
we should expect people to rebel if and only if the post-revolutionary benefits exceed all the
revolutionary switching costs — including, above all, a substantial risk of death.
I have previously argued that Mises’ analysis of democracy was well ahead of its time
(Caplan and Stringham 2005). It is hard to say the same about his analysis of dictatorship.
To claim that dictators must obey public opinion to avoid violent overthrow is a gross
exaggeration. To add the qualifier “in the long run” is only a small improvement. Tullock’s
paradox of revolution and switching costs plainly leave dictators with considerable slack.
Nevertheless, Mises has good political intuition. Unlike most social scientists, he does not
expect foolish policies to be unpopular. Therefore, when dictatorships impose foolish policies,
Mises is unusually open to the possibility that their policies could have passed the democratic test.
Mises’ key mistake is to hastily infer that dictators must choose popular policies to avoid
revolution. There are a number of other mechanisms that could explain why dictators and
their policies would usually be popular. Perhaps dictators have a great deal of slack but
decline to use it. Choosing unpopular policies might hurt economic growth, which benefits
dictators and their subjects alike. Dictators might prefer popular policies because they share
the values and beliefs of their subjects. More strikingly, however, dictators may be able to
shape public opinion and policy. If human beings suffer from status quo bias, dictators can
make a wide range of policies popular simply by adopting them. No matter which
mechanism matters most, however, Mises’ Democracy–Dictatorship Equivalence Theorem
underestimates the power of the dictator.
Acknowledgment I thank Geoffrey Lea for excellent research assistance, Pete Boettke and Tyler Cowen for
many helpful discussions, David Gordon, Alex Tabarrok, and an anonymous referee for detailed comments,
and the Mercatus Center for financial support.
58
B. Caplan
References
Alesina, A., & Fuchs-Schündeln, N. (2005). Good bye Lenin (or not?): The effect of communism on people’s
preferences. NBER working paper 11700.
Applebaum, A. (1994). The fall and rise of the communists: Guess who’s running central Europe? Foreign
Affairs, 73(6), 7–13.
Barro, R. (1997). Determinants of economic growth: A cross-country empirical study. Cambridge, MA: MIT
Press.
Caplan, B. (2007). The myth of the rational voter: Why democracies choose bad policies. Princeton, NJ:
Princeton University Press.
Caplan, B., & Stringham, E. (2005). Mises, Bastiat, public opinion, and public choice: What’s wrong with
democracy. Review of Political Economy, 17(1), 79–105.
Colton, T. (2000). Transitional citizens: Voters and what influences them in the New Russia. Cambridge:
Harvard University Press.
Conquest, R. (1986). Harvest of sorrow: Soviet collectivization and the terror-famine. New York: Oxford
University Press.
de la Boetie, E. (1975). The politics of obedience: the discourse of voluntary servitude. New York: Free Life
Editions.
Deacon, R., & Saha, S. (2006). In A. Ott, & R. Cebula (Eds.) Public good provision in dictatorships: A
survey. The Elgar companion to public economics: Empirical public economics. Elgar: Cheltenham, UK.
Fessenden, H. (2002). Red dawn. Foreign Affairs, 81(1), 177–91.
Fu, Z. (1993). Autocratic tradition and chinese politics. Cambridge: Cambridge University Press.
Galetovic, A., & Sanhueza, R. (2000). Citizens, autocrats, and plotters: A model and new evidence on coups
d’etat. Economics and Politics, 12(2), 183–204.
Grossman, H. (2000). The state: Agent or proprietor? Economics of Governance, 1(1), 3–11.
Gryzmała-Busse, A. (2002). Redeeming the communist past: The regeneration of communist parties in East
Central Europe. Cambridge: Cambridge University Press.
Höijer, R. (2004). Theft by bandits and taxation by kings: A critique of Mancur Olson on state formation.
Political Studies Review, 2(1), 24–38.
Hummel, J. (1990). National goods versus public goods: Defense, disarmament, and free riders. Review of
Austrian Economics, 4(1), 88–122.
Jones, B., & Olken, B. (2005). Do leaders matter? National leadership and growth since World War II.
Quarterly Journal of Economics, 120(3), 835–64.
Klemperer, P. (1995). Competition when consumers have switching costs: An overview with applications to
industrial organization, macroeconomics, and international trade. Review of Economic Studies, 62(4),
515–39.
Kuran, T. (1995). Private truths, public lies: The social consequences of preference falsification. Cambridge:
Harvard University Press.
Kurrild-Klitgaard, P. (1997). Rational choice, collective action and the paradox of rebellion. Copenhagen,
Denmark: Institute of Political Science and Political Studies Press.
Lewis, P. (2001). Party development and democratic change in post-communist europe: The first decade.
London: Frank Cass.
Malia, M. (1994). The soviet tragedy: A history of socialism in Russia, 1917–1991. New York: The Free
Press.
McGuire, M., & Olson, M. (1996). The economics of autocracy and majority rule: The invisible hand and the
use of force. Journal of Economic Literature, 34(1), 72–96.
Mill, J. S. (2006). Principles of political economy, with some of their applications to social philosophy. In
J. S. Mill (Ed.) Collected works of John Stuart Mill, Volume III. Indianapolis, IN: Liberty Fund.
Mises, L. v. (1978). Liberalism: A socio-economic exposition. Kansas City: Sheed, Andrews, and McMeel.
Mises, L. v. (1980). Planning for freedom. South Holland, IL: Libertarian Press.
Mises, L. v. (1981). Socialism: An economic and sociological analysis. Indianapolis, IN: Liberty Fund, Inc.
Mises, L. v. (1983a). Bureaucracy. NY: Arlington House.
Mises, L. v. (1983b). Nation, state, and economy. NY: New York University Press.
Mises, L. v. (1985a). Theory and history. Auburn, AL: Mises Institute.
Mises, L. v. (1985b). Omnipotent government. Spring Mills, PA: Libertarian Press, Inc.
Mises, L. v. (1998a). Human action. Auburn, AL: Mises Institute.
Mises, L. v. (1998b). Interventionism: An economic analysis. Irvington-on Hudson, NY: Foundation for
Economic Education.
Mosca, G. (1939). The ruling class. New York: McGraw-Hill.
Mises’ democracy–dictatorship equivalence theorem: A critique
59
Mulligan, C., Gil, R., & Sala-i-Martin, X. (2004). Do democracies have different public policies than
nondemocracies? Journal of Economic Perspectives, 18(1), 51–74.
Olson, M. (2000). Power and prosperity. New York: Basic Books.
Pinker, S. (2002). The blank slate: The modern denial of human nature. New York: Viking.
Pipes, R. (1974). Russia under the old regime. New York: Charles Scribner’s Sons.
Pipes, R. (1990). The Russian revolution. New York: Vintage Books.
Pipes, R. (2004). Flight from freedom: What Russians think and want. Foreign Affairs, 83(3), 9–15.
Przeworski, A., Alvarez, M., Cheibub, J., & Limongi, F. (2000). Democracy and development: Political
institutions and well-being in the world 1950–1990. New York: Cambridge University Press.
Salisburg, H. (1992). The new emperors: China in the era of Mao and Deng. New York: Avon.
Samuelson, W., & Zeckhauser, R. (1988). Status quo bias in decision making. Journal of Risk and
Uncertainty, 1(1), 7–59.
Schelling, T. (1978). Micromotives and macrobehavior. New York: Norton.
Shapiro, C., & Varian, H. (1999). Information rules: A strategic guide to the network economy. Boston:
Harvard Business School Press.
Shikaki, K. (2004). The future of Palestine. Foreign Affairs, 83(6), 45–60.
Tullock, G. (1974). The social dilemma: The economics of war and revolution. Blacksburg, VA: University
Publications.
Tullock, G. (1987). Autocracy. Dordrecht, The Netherlands: Kluwer.
Velasco, A. (1994). The state and economic policy: Chile 1952–92. In B. Bosworth, R. Dornbusch, & R. Labán
(Eds.) The Chilean economy: Policy lessons and challenges (pp. 379–429). Washington, DC: Brookings
Institution.
Werth, N. (1999). A state against its people: Violence, repression, and terror in the Soviet Union. In S. Courtois,
N. Werth, J.-L. Panne, A. Paczkowski, K. Bartošek, & J.-L. Margolin (Eds.) The black book of
communism: Crimes, terror, repression (pp. 32–268). Cambridge: Harvard University Press.
Wintrobe, R. (1998). The political economy of dictatorship. Cambridge: Cambridge University Press.